Miniprogram Skills - Reusable Codex & Claude Code Skills for WeChat Mini Program Development

English | 中文

Reusable AI coding skills for WeChat Mini Program development, DevTools diagnosis, scaffold validation, and miniapp architecture optimization. Works with Codex and Claude Code.

What is Miniprogram Skills?

Miniprogram Skills is an open-source AI coding skill library specifically designed for WeChat Mini Program development. It provides 6 reusable skills that help developers diagnose DevTools issues, validate project scaffolds, optimize miniapp architecture, and improve user-facing copy.

The project supports 2 AI coding platforms (Codex and Claude Code) and includes complete bilingual documentation in English and Chinese.

Key Statistics

Metric	Value
Reusable Skills	6
AI Platforms Supported	2 (Codex, Claude Code)
GitHub Topics	13
Documentation Languages	2 (English, Chinese)
Evaluation Fixtures	10+
Machine-Readable Catalog	1 (JSON)

Why This Exists

Based on WeChat Mini Program Official Documentation and WeChat DevTools Documentation, common issues include:

• WeChat Mini Program setup and DevTools failures repeat across projects
• CLI-visible failures and GUI-only failures need different workflows
• Wrong-root imports, stale compile settings, and TypeScript recognition drift are common
• These patterns are more reusable than any one product repo

Features

• 6 Reusable Skills - Covering scaffold validation, DevTools recovery, CLI diagnosis, GUI checks, architecture refactoring, and copy optimization
• Dual Platform Support - Works with both Codex and Claude Code
• TypeScript Ready - Full TypeScript support for miniapp development
• Bilingual Documentation - Complete English and Chinese documentation
• Machine-Readable Catalog - JSON-based skill catalog for automation
• Routing Evaluation - Replayable fixtures for skill boundary testing

Current Public Skills

Skill	Purpose	Status
miniapp-official-scaffold-alignment	Validate or design an officially valid miniapp scaffold	public draft
miniapp-devtools-recovery	Recover a repo after wrong-root import, template residue, or compile drift	public draft
miniapp-devtools-cli-repair	Diagnose DevTools through the official CLI and classify safe repo-level fixes	public draft
miniapp-devtools-gui-check	Run or design host-side GUI smoke checks for runtime-only failures	public beta tool
miniapp-center-hub-refactor	Reorganize a growing miniapp into a clearer center or hub with stable ownership	public draft
miniapp-user-facing-copy-trim	Shorten verbose miniapp copy into clearer action-first labels and summaries	public draft

Use Case Scenarios

Scenario	Recommended Skill	Description
DevTools shows errors	miniapp-devtools-cli-repair	Diagnose CLI-visible failures through official CLI
Wrong root import	miniapp-devtools-recovery	Recover from wrong-root imports or template residue
Scaffold validation	miniapp-official-scaffold-alignment	Validate project structure against official guidelines
GUI runtime issues	miniapp-devtools-gui-check	Check GUI-only runtime failures
Navigation reorganization	miniapp-center-hub-refactor	Reorganize scattered tabs into hub structure
Verbose UI copy	miniapp-user-facing-copy-trim	Simplify user-facing text

Quick Start

For Codex Users

Skills are automatically loaded from the skills/ directory. Each skill has an agents/openai.yaml configuration file.

For Claude Code Users

Skills can be triggered in two ways:

1. Auto-trigger: Claude Code automatically invokes skills based on the description field
2. Manual trigger: Use slash commands like /miniapp-devtools-recovery

Installation

# Clone the repository
git clone https://github.com/Sun-sunshine06/miniprogram-skills.git

# Navigate to the project
cd miniprogram-skills

# Run validation
pwsh -File scripts/check.ps1

Repository Layout

.
├── docs/                           # Documentation
│   ├── conventions.md              # Writing conventions
│   ├── public-roadmap.md           # Development roadmap
│   └── skill-map.md               # Skill overview
├── evals/                          # Evaluation fixtures
│   ├── routing/                    # Routing test cases
│   └── routing-replays/           # Replay transcripts
├── manifests/                      # Machine-readable catalogs
│   └── skill-catalog.json         # Public skill catalog
├── schemas/                        # JSON schemas
├── skills/                         # Skill definitions
│   ├── miniapp-devtools-cli-repair/
│   ├── miniapp-devtools-gui-check/
│   ├── miniapp-devtools-recovery/
│   ├── miniapp-center-hub-refactor/
│   ├── miniapp-user-facing-copy-trim/
│   └── miniapp-official-scaffold-alignment/
└── tools/                          # Utilities
    └── wechat-gui-check/          # GUI smoke checker

Platform Compatibility

All skills support both Codex and Claude Code:

• Codex: Reads name and description from SKILL.md, uses agents/openai.yaml for platform config
• Claude Code: Reads all SKILL.md fields including optional tools and slash_command

Slash Commands

Each skill can be triggered manually in Claude Code using its slash command:

Skill	Slash Command
miniapp-official-scaffold-alignment	/miniapp-official-scaffold-alignment
miniapp-devtools-recovery	/miniapp-devtools-recovery
miniapp-devtools-cli-repair	/miniapp-devtools-cli-repair
miniapp-devtools-gui-check	/miniapp-devtools-gui-check
miniapp-center-hub-refactor	/miniapp-center-hub-refactor
miniapp-user-facing-copy-trim	/miniapp-user-facing-copy-trim

Frequently Asked Questions (FAQ)

What is Miniprogram Skills used for?

Miniprogram Skills is an open-source AI coding skill library for WeChat Mini Program development. It provides reusable skills for diagnosing DevTools issues, validating project scaffolds, optimizing miniapp architecture, and improving user-facing copy.

How do I use these skills with Claude Code?

You can use skills in Claude Code in two ways:

1. Auto-trigger: Claude Code automatically invokes skills based on the description field in SKILL.md
2. Manual trigger: Use slash commands like /miniapp-devtools-recovery or /miniapp-official-scaffold-alignment

Which AI coding platforms are supported?

Miniprogram Skills supports 2 AI coding platforms:

• Codex (OpenAI) - Uses agents/openai.yaml configuration
• Claude Code (Anthropic) - Uses SKILL.md frontmatter with tools and slash_command fields

What types of issues can these skills diagnose?

The skills can diagnose:

• CLI-visible failures: Compile errors, preview issues, service port problems
• GUI-only runtime issues: Page load failures, interaction problems, WebSocket issues
• Project structure issues: Wrong root imports, template residue, TypeScript recognition drift
• Architecture problems: Scattered navigation, verbose UI copy

Is this project actively maintained?

Yes, the project is actively maintained with regular updates. The current version is v0.4.0 with 6 public skills and comprehensive evaluation fixtures.

How do I contribute to this project?

See CONTRIBUTING.md for guidelines. The project welcomes contributions including new skills, bug fixes, documentation improvements, and forward-testing on different miniapp repos.

Local Validation

Run the shared local validation flow before opening a PR:

pwsh -File scripts/check.ps1

This command validates:

• Public skills structure and metadata
• Machine-readable skill catalog
• Routing fixtures and replay transcripts
• Negative-path fixtures
• Markdown links and bilingual cross-links
• Repository JSON validity
• Tool syntax
• External-project dry-run smoke checks

Contributing

See CONTRIBUTING.md for guidelines.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Links

• Documentation
• Skill Map
• Public Roadmap
• Release Notes
• Changelog